The response of structures to dynamic crowd loads

Digest 426
2004 Edition

digest

B R Ellis T Ji

BRE Centre for Structural and Geotechnical Engineering

Manchester Centre for Civil and Construction Engineering, UMIST

The 1996 edition of the British Standard BS 6399-1, ‘Loading for buildings, Code of practice for dead and imposed loads’, included guidance on dynamic loads generated by synchronized crowd movement. This Digest provides information that explains and supports the recommendations in the Standard, and provides a method for calculating structural response to rhythmic crowd loads. This edition of Digest 426 supersedes the version published in 1997. It includes improved information on the loads generated by crowds.

Dynamic crowd loads are generated by the movement of people. However the largest loads are produced by synchronized rhythmic movements which mainly arise from people dancing or engaging in jumping type movements, usually in response to a musical beat. An example of a dance which involved jumping, and which was popular in the 1980s, was ‘pogoing’. Although usually associated with lower numbers of people, activities such as aerobics will give rise to similar dynamic effects. It should be recognised that a crowd of people jumping rhythmically can generate large loads and this may be a concern for both safety and serviceability evaluations. However this is an extreme case and the rhythmic excitation most frequently encountered does not involve everyone jumping in perfect synchronization but includes people dancing and clapping, often with some people stationary. This latter situation will produce smaller loads than if everyone was

jumping and usually a much smaller structural response. For design it is therefore important to identify an appropriate load scenario. Two approaches to the design of structures to accommodate dynamic crowd loads are given in British Standards and this Digest. One relies on ensuring that the fundamental natural frequency of the structural system is sufficiently high that resonance generated by the major dynamic components of the load is avoided; the other provides a method for calculating structural response to dynamic crowd loads so that structural safety and serviceability limit states can be checked. This Digest deals primarily with the modelling of dynamic crowd loads and the calculation of structural response to these loads. It gives the background to the problem and discusses a number of items which must be considered in design.

arises when a coordinated jumping load induces a resonant response in a structure. it is only when coordination is achieved that the load becomes significant.. no mention was made of dynamic loads generated by crowds. for one type of jumping. it now refers to the jumping loads through reference to the 1997 edition of Digest 426 where the use of the load model can be explained.. if a group of people on a structure were jumping in an uncoordinated manner. ‘Where significant dynamic loads are to be expected the structure should be designed either: a) to withstand the anticipated dynamic loads. a resonant situation would arise whereby the response would be amplified greatly. A simple illustration The history of dynamic crowd loads and their inclusion in UK Standards From experience. will jump on it. the response of a floor to 64 people jumping on it in time to a musical beat was monitored. resonance will occur which can greatly amplify its response. at that time there was no authoritative guidance on acceptable serviceability levels for this type of vibration. or b) by avoiding significant resonance effects ‘To avoid resonance effects the vertical frequency should be greater than 8. The guidance considers the loads generated by the activity which are independent of the structure to which they are applied. If the synchronized movement excites a natural frequency of the affected part of the structure. It has been mentioned that BS 6399-1:1996 provides guidance on loads arising from synchronized crowd movement. jumping. If the frequency of the dynamic load occurs at a natural frequency of the structure. With structures which are designed for static loading. if a person keeps jumping the resulting dynamic load will be cyclic. the peak load is 4. The dynamic displacements were 3.. allow for small dynamic effects . The loads do not . but did not exclude the use of other load models. Although everyone jumping on a dance floor may be possible. static design loads were detailed but with the warning ‘the values for imposed loads given . it was questioned whether this model was suitable for grandstands. In the 1967 version of BS CP3. but for crowds the dynamic loads will only be significant when the movement is synchronized. . This is demonstrated by the action of a child who.. Hence at this stage the problem had been recognised although no guidance was provided on how to deal with it. The real problem. For example. This provides an opportunity for specialist documents to be available for specific types of construction and use (eg grandstands). BS 6399-1. It should be noted that the above frequency recommendations relate to safety assessments and not serviceability. In the 1984 edition of BS 6399 Part 1. rhythmic stamping. The Standard. Unfortunately this resulted in the majority of engineers ignoring the problem... so at this stage the problem cannot have been considered to be important. aerobics etc) is synchronized. There are some human activities which involve repeated jumping. Guidance on dynamic crowd loads was included in BS 6399-1:1996 specifically to take account of loads generated by repetitive jumping during dancing. In October 2002. In fact. allow for dynamic loads due to crowds’. BS 6399-1 was amended. Indeed. then the displacement due to resonance would be 25 times the static displacement. hence in certain situations it can generate a resonant response of the loaded structure. however. It is instructive to provide some quantitative values to illustrate the magnitude of the problem. the frequencies being evaluated for the appropriate mode of vibration of an empty structure. Later in the Digest an equation is given which determines the load produced by one person jumping rhythmically. gave one example of a load model for everyone in a crowd jumping. It is perhaps wise to recognise that structural resonance produced by dynamic crowd loads will be an unusual event and requires both a specific form of loading and a structure which is vulnerable to such dynamic loads.4 Hz and the horizontal frequencies greater than 4. Therefore this section provides a simple illustration of the problem and records related developments in UK Standards. Furthermore. when trying to crush a cardboard box. In practice this only occurs in conjunction with a strong musical beat such as occurs at lively pop concerts or aerobics. Such crowd movement can generate both horizontal and vertical loads. While it retains the frequency limits for synchronized excitation. and is therefore a situation to be considered.’ To withstand the anticipated dynamic loads reference is made to published work [1]. It is of interest to examine the two versions of the British Standard which preceded it. everyone is aware that the peak load produced by jumping is significantly larger than the load resulting from standing still.7 times the person’s weight. the peak displacement induced by the load would not be significantly larger than that induced by the static load. there are many safety factors built into the design which allow for most dynamic loads. The dynamic loading is thus related to the dance frequency or the beat frequency of the music and is periodical.5 times the static displacement for the group.2 Background It is possible that some designers may be unaware of the nature of the potential problems caused by dynamic crowd loads. if a structure with a damping value of 2% critical is subjected to a dynamic load coinciding exactly with the natural frequency of the structure. In practice this usually occurs in conjunction with a strong musical beat at events like lively pop concerts and aerobic exercise classes. It should be appreciated that this is a significant load case. In an experiment which is described later. Hence it can be appreciated that a dynamic load occurring at a natural frequency of a structure can induce significant vibrations.0 Hz. The relevant section of the Standard states: ‘Dynamic loads will only be significant when any crowd movement (dancing.

as already mentioned. Second. those furthest from the musical source will hear the music slightly later than those close to the source. or to individuals dancing with different styles or enthusiasm. Crowd size Crowd size and the area a crowd occupies are also important.5 Hz for new grandstands for normal. The logical development. It means that the loads produced by a group are not a simple summation of the ideal loads produced by individuals. consequently if the loading function is expressed in Fourier series. for a crowd. but it was recognised that these limits are not a perfect design solution and many items affect a grandstand’s performance besides its natural frequency. This is because the loading is cyclic but is not a simple sinusoid. Therefore actual static loads appropriate to the activity should be used in determining dynamic loads. because there is always an opportunity for a normal crowd to start jumping rhythmically. the response to any given load. several factors relating to the crowd. Second. a stationary crowd acts as an additional mass–spring–damper system on a structure and should be modelled accordingly[7]. Consequently a joint working group of the Institution of Structural Engineers. the Department for Culture Media and Sport. crowd activity and the structure need to be considered. and the Office of the Deputy Prime Minister was set up to consider dynamic loading of seating decks. Therefore for evaluation of structural response to dance-type loads. The frequency limits were based on the experience of the committee members. a static design load of 5 kN/m2 is used.5 Hz. This is considered in greater detail by Ginty et al[5] and Littler[6].5 to 3. Dynamic crowd effect If a crowd of people all tried to jump at the same frequency. alarming vibrations have been observed at three major football stadia in England following rhythmic excitation[3]. synchronized crowd movement is usually coordinated by music. Human–structure interaction General considerations To define crowd loads. For the situation where many in the crowd are moving rhythmically but a significant number are stationary. One of the difficulties of modelling structures with crowds of people is how to account for the mass of people since it can affect the structural characteristics and. which may be more common at the extremes of the frequency range. is to define appropriate load models for an individual and a crowd.8 is more realistic. This is considered later. This equates roughly to six people per square metre – a relatively dense crowd. Crowd density With many structures which are subject to crowd loads. the cyclic nature of the activity constrains the Fourier components to be integer multiples of the dance frequency. the higher frequency jumping cannot be sustained and an upper limit of 2. As an interim measure the working group suggested limits for vertical frequencies of 3. Imperfect coordination can be due either to individuals not jumping at exactly the beat frequency. Even the distinction between stands subject to normal and rhythmic loading is somewhat contentious. The mode shapes of bays in the structure often show the motion in adjacent bays as being anti-phase. which may be the case for some dances. the characteristics of the unloaded structure should be used in the calculations. consequently. In fact there is a range of different situations of which the two extremes are important. thereby affecting the overall coordination – a phenomenon which has been observed in a grandstand[4]. These observations resulted in corrective measures being introduced to the stadia. possibly due to seating or licensing requirements. With people dancing or exercising which involves repeated jumping. the stationary people will provide a significant damping mechanism and a consequent reduction in any resonant response. or to phase differences between individuals. First. In fact there is an attenuation of the peak load with increasing group size which is greater for the higher Fourier components. and may well be found with some dance floors. First. These models will then enable responses to be calculated and specific structural configurations to be evaluated. the structure required to accommodate a large group is likely to consist of many different structural bays and areas. non-rhythmic loading but 6 Hz for those where pop concerts could be held[2]. but. This will induce a phase difference in response with distance from the musical source.3 The response of grandstands to dynamic crowd loads became the focus of attention towards the end of the 1990s – in particular the response of the cantilever tiers of some grandstands. hence loading in one bay may effectively attenuate the response in other bays. . Frequency range The frequency range for an individual person jumping is approximately 1. for the case where people are moving (eg walking and jumping) their body mass is not involved in the vibration of the structure and the human involvement is simply as a load. their coordination will not be perfect. in many instances the number of people in a given area is limited. However. the energy is not only generated at the dance frequency but is produced at whole number multiples of the dance frequency. For a large crowd. The critical situation is often reached by the loading in a single bay. starting from the simple limiting frequency.

2 0. Indeed. By equating the mean value of Equation 1 over one period Tp to the weight Gs the following relationship is obtained: K= π 2α (Equation 3) { 0 ≤ t ≤ tp tp ≤ t ≤ Tp (Equation 1) where: Kp = the impact factor (Fmax /Gs) Fmax = the peak dynamic load Gs = the weight of the jumper t p = the contact duration Tp = the period of the jumping load (N) (N) (s) (s) Thus the loads can be determined knowing the weight of the jumper.4 Loading Load model for an individual jumping The load–time history for continuous jumping can be described by a high contact force for a certain time tp (contact period) followed by zero force when the feet leave the floor.4 Time (s) 0. α 2 1 1 /3 /2 /3 5 Equations 5 & 6 (n = 6) Equation 1 4 α = 1/3 f = 2 Hz Normalized load 3 2 1 0 –1 0. Figure 1 shows a normalized load–time history for jumping with α = 1/3 and fp = 1/Tp = 2. The load function in one period for a single person Fs(t) is given by: Fs(t) = Kp Gs sin (π/tp) 0 It has been observed experimentally[9] that the mean value of the time history of a vertical load corresponding to bouncing to music on toes (jouncing) or to rhythmic jumping is always equal to the weight of the performer. the period of the jumping and the contact ratio.0 corresponds to the static weight of the person.7 times the static load.0 Figure 1 Load–time history for one person jumping . this aligns well with measurements for individuals jumping.0 Hz.0 0. the lower the contact ratio and the higher the peak dynamic load. The contact ratio α is defined as follows: α= tp ≤ 1. From the figure it can be appreciated that the dynamic load can be significantly larger than the static load.0 Tp (Equation 2) Table 1 Typical values of the contact ratio α for various activities[8] Activity Low impact aerobics Rhythmic exercises. which can be seen in Table 1. It can also be deduced that the higher the jumping.6 0. It has been proposed that the load–time function can be expressed by a sequence of semi-sinusoidal pulses. The normalized load of 1.8 1. high impact aerobics Normal jumping Therefore different contact ratios characterise different rhythmic activities. The contact period tp can vary from 0 to Tp corresponding to different movements and activities. from Equation 3 the peak dynamic load can be calculated directly which for these parameters is 4.

61 x v – 0. Hence for each Fourier term the average load per cycle is zero.v = 0. The FCs were plotted against the number of people in the group on a log/log plot together with regression lines for each FC based on a power relationship.t) = distributed force which varies with time G (x. For the experiments only the lowest three Fourier coefficients (FCs) were determined. For calculation purposes both the Fourier coefficients and the phase angles are required. calculated using Equations 5 and 6.v = nth Fourier (load) coefficient induced by v persons Here it is necessary to consider the spatial distribution of the people. Using the FCs defined above and the theoretical phase angles for normal jumping (Table 2.082 ● first FC r2. This has been determined experimentally and reported in detail in IP 4/02.v = 0. tests were undertaken on two floors with different sized groups jumping at selected frequencies. α = 1/3). This aligns with the observation upon which Equation 3 was derived. In the Equation the terms within the summation sign are the Fourier components of the load and occur at multiples of the load frequency. the values for the load factors and phase lags can be determined for jumping using the following equations (Ji and Ellis[1]. This can be extended to cover crowds.y. is shown in Figure 1 along with the equivalent load determined using Equation 1. Table 2 lists the first six Fourier coefficients and phase lags for different contact ratios.0 + ∑ rn. Also if the dancing stopped the dynamic loads would become zero and the resulting force would simply be the weight of the dancer.24 ● second FC r3. The group sizes were selected to be powers of 2 with 64 being the largest group that could be accommodated safely.5 Load model for an individual in terms of Fourier series Load models for crowds jumping: experimental values For analysis purposes it is useful to express Equation 1 in terms of Fourier series: 2nπ Fs (t) = Gs 1.t) = G (x. and including the first six Fourier terms. Based upon Equation 3. In the experiments. The jumping was coordinated using a musical beat at selected frequencies.v sin t + φn n =1  Tp  where: rn = Fourier coefficient (or dynamic load factor) of the nth term n = number of Fourier terms φn = phase lag of the nth term Equation 4 can represent the vertical load for an individual for any form of rhythmic movement. The load equation for a crowd jumping at a comfortable frequency becomes: 2nπ F (x.0 + ∑ rn sin t + φn n =1  Tp  { } (Equation 4) In the previous sections the loads produced by individuals have been considered. However there is also a variation in the values of the Fourier coefficients due to the different jumping styles and coordination of the jumpers.y) = density and distribution of human loads rn.v = 1. Table 2 Fourier coefficients and phase lags for different contact ratios n=1 n=2 n=3 n=4 n=5 n=6 α = 2/3 α = 1/2 α = 1/3 rn φn rn φn rn φn – π/6 π/2 0 9/ 5 9/ 7 9/ 55 –5 π/6 2/ 15 – π/2 0 9/ 247 – π/6 π/2 9/ 391 –5 π/6 0 0 2/ 63 – π/2 π/2 π/2 2/ 3 – π/2 π/6 2/ 15 – 2/ 35 – π/6 9/ 7 – 0 2/ 3 – π/2 9/ 55 –5 π/6 9/ 91 – π/6 2/ 15 – . The data are shown in Figure 2 on page 6.y) 1. and Ji and Wang[10]): { } (Equation 7) where: F (x. the load model predicted similar amplitudes of acceleration and displacement to the average measured values for the various group sizes up to 64 people.44 x v – 0. rn = { π/2 2cos (nπα) 1 – (2nα)2 0 if 2nα = 1 if 2nα ≠ 1 (Equation 5) φn =   if 2nα = 1  1 + cos(2nπα) sin(2nπα) tan–1 – π if <0  sin (2nπα)  1 – (2nα)2 π – 2 1 + cos(2nπα) tan–1  sin(2nπα)  if sin(2nπα) =0 if sin(2nπα) >0 1 – (2nα)2   if 2nα ≠ 1  (Equation 6) The normalized load–time history. it can be seen that the FCs fall between the values given in Table 2 for normal jumping (α = 1/3) and rhythmic exercises (α = 1/2). The regression lines are defined by the following equations which characterise the variation in FCs with the number of people v in the group: r1.31 ● third FC If only one person is considered (ie v = 1).y.94 x v – 0.

it should be recognised that these are still significant loads. To demonstrate this point.6 3 2 10 0.86 kN/m2.0 9 8 7 6 First Fourier coefficients Second Fourier coefficients Third Fourier coefficients 10 0 2 3 4 5 6 7 8 9 101 2 3 4 5 6 7 8 Number of people Figure 2 Fourier coefficients determined for all of the tests 4 Level for unloaded floor 2 Displacement (mm) 0 –2 –4 –6 0 2 4 Time (s) 6 8 Figure 3 Measured displacements for 64 people on the second floor Although these FCs reduce with increasing crowd size.185 people/m2 which is not densely packed. The floor area was 9 x 6 m. Taking the average weight of the jumpers gives a distributed load of 0. hence 64 people represent a density of 1. The measured displacement (static + dynamic) was actually 3. This is shown in Figure 3. it is worth examining the central displacement of the floor measured for 64 people jumping.47 times the static displacement .0 9 8 7 6 Fourier coefficient 5 4 3 2 10 –1.

then the calculated dynamic displacement–time history for the second floor is as shown in Figure 4. gives an equivalent peak dynamic load of 2.99 kN/m2. The floor was designed for 3.7 4 2 Displacement (mm) 0 –2 –4 –6 0 2 4 Time (s) 6 8 Figure 4 Calculated displacements for 64 people on the second floor 2 10 0. . This load scenario did not include a significant resonant component. If the FCs for a group of 64 people (v = 64) are used in Equation 7 to define the load.0 9 8 7 6 First Fourier coefficients Second Fourier coefficients Third Fourier coefficients 100 2 3 4 5 6 7 8 101 2 3 4 5 6 7 8 102 2 3 4 5 6 7 8 103 2 3 4 5 6 7 8 104 Number of people Figure 5 Fourier coefficients against group size: combined model which.0 9 8 7 6 Fourier coefficient 5 4 3 2 10 –1.5 kN/m2. when used to factor the static load.

42. The distribution of variables was derived from the experimental measurements. the first three FCs of the loads were 0. Based upon the assumption that the FCs varied with the number of people in a similar fashion to the jumping tests. It should be remembered that these calculations are based upon experiments with students undertaking the jumping. whereas in other groups the coordination may be far worse. A structure which may perform adequately during concerts may appear to be unacceptable when subjected to jumping loads. dancing and moving in response to music but not jumping. It is likely that some groups of people (eg professional dancers) will be better coordinated than the students (and therefore produce higher peak loads).61.44).94 and 0. the situation with everyone jumping does not arise.017. As everyone was not moving the damping of the occupied structure is increased significantly from that of the bare structure due to human–structure interaction. it was seen that the spread in calculated FCs became less for the larger groups. This common scenario leads to lower peak loads than those produced by everyone jumping and a significantly smaller structural response. 0. With the results for up to 64 people based upon six calculations for each group size. Vertical loads for synchronized movement not involving jumping The previous sections have considered rhythmic jumping which. Further evaluations are given by Ellis and Ji[11].192 people were 1. the variation in loads with group size is seen in Figure 5 on page 7. can generate high loads. for the final FCs there was no longer a reduction in FC with crowd size. ● the phase difference between the prompt and the individual jumping. in many situations involving rhythmic excitation (eg pop concerts). This was probably due to the statistical variations being smoothed by a larger number of samples. The load model for an individual jumping (Equation 4) was used as the basis for the numerical modelling with the aforementioned variables introduced into this model in order to calculate a load–time history. ● the jumping frequency f which may not be perfectly aligned with the prompt. it is appreciated. The values for the first three FCs for groups of 8. although this required several assumptions since the loading on the grandstands was not artificially controlled[12]. For an individual jumping in response to an audible prompt at a given frequency. However. In the 1990s. It can be appreciated that these are much smaller than those for jumping (ie 1.072. were used to examine the effects of larger group sizes. Analysis of recorded data concentrated on the largest responses which were when approximately 90% of the audience was standing. 0. and which correlate well with the experiments described in the previous section. A damping value of 16% was determined from the analysis of the measurements which would have a significant effect upon any resonant response.8 Load models for crowds jumping: numerical values Although the experiments provide some useful results they are limited to a relatively small group of people. . There is one further important factor. Usually some people will be jumping or dancing and others standing or sitting. the following variables were considered: ● the contact ratio α and therefore the height of jumping. BRE monitored a number of events in cantilever grandstands during concerts. suggesting that the power relationships derived from the experimental data are only appropriate for the smaller groups. Groups of people were represented by the combination of the appropriate number of individual load–time histories. One example will be given to illustrate this. The results which were obtained. Furthermore. Similar methods to those developed for the jumping tests were used to determine the loads.40. hence the experiments were modelled numerically.47 and 0.087 and 0. 0. Therefore it should be recognised that it is important to define the appropriate loading scenarios for design. with the objective of replicating the variation in loads observed in the experiments.

and A(t) is the amplitude of vibration corresponding to that mode and is a function of time. that should be considered and not necessarily the fundamental mode of the whole grandstand. for serviceability assessments. the value of B would be 1. This ensures that any resonant excitation of the fundamental mode will be considered. However. for other structures it may be another mode of vibration which is critical. the first I Fourier terms should be included in the determination of acceleration where I is defined as the first integer bigger than fs /f. Structural response For this section it is assumed that the structural mass and . and the shape of this mode is relatively easy to choose with sufficient accuracy for many common cases. For example. it is not necessary to use all the terms in the Fourier series to describe the load. Basically. only the response from the fundamental mode needs to be considered. If the correct mode shape was used to determine B for a rectangular plate with fully fixed boundary conditions.t) = A(t) Φ(x. if a cantilevered tier of a grandstand is being examined. Also the modal load for the second symmetric mode will be significantly lower which effectively means that the response will be dominated by the contributions of the fundamental mode.y. Alternatively the loads could be used in a finite element programme to determine the response of a structure. it is recommended that the first three harmonic terms should be considered. The structural factor The structural factor B relates to the fundamental mode and depends on the type of structure and boundary conditions. It should be noted that resonance can occur when the natural frequency of the structure fs is equal to the dance frequency f or to integer multiples of the dance frequency. For a single bay floor with symmetric boundary conditions. only the fundamental mode needs to be considered. For safety assessments. for a single bay floor. under a uniformly distributed load. this is termed the principal mode for that location. this is one situation that should be evaluated. The number of Fourier terms to be taken into account depends on their contribution to the response. Selection of the appropriate mode of vibration For calculation purposes. Therefore. the structural factor can. The solutions provided by Ji and Ellis[1] can be applied to structures with different boundary conditions. These calculations include all the modes of vibration of the system and all the Fourier components of the load. there will be no vibration of the antisymmetric modes. including any possible resonance. it would be the vertical mode of vibration of the cantilever.9 Structural response Initial considerations Number of Fourier components to consider Given the number of people involved and their spatial distribution. This provides formulae for calculating how displacement and acceleration vary with time. be defined as follows: B= ∫∫ Φ(x. If both the structural mass and the loading are uniformly distributed in space. according to the solution procedure.y) dxdy s (Equation 9) For a simply supported rectangular plate the mode shape would be: Φ(x. and the frequency and type of the coordinated movement. As resonant excitation may produce the largest response. The resulting value of B would equal 1. The response of a structure can be approximated by the contribution of the fundamental mode: w(x. no anti-symmetric modes are involved in the vibration. or part of the cantilever.y) is the dimensionless fundamental mode with unit peak value. Consequently it is the appropriate mode or modes of vibration for the analysis that must be determined.72. General solution For many structures it will be the fundamental mode of vibration which will be important and this term will be used in the rest of this Digest.y) = sin (πx /Lx) sin (πy /Ly) (Equation 10) where Lx and Ly are the length and width of the plate.62. The difficulty encountered is in choosing suitable displacement functions.y) dxdy s ∫∫ Φ2 (x. will be the fundamental mode. This deflected shape may be reasonable for many floors. Considering a symmetric floor in a single bay under a symmetric load. Therefore. the overall load can be determined. where evaluation of displacement and stresses is appropriate. For a more complicated structure it may be necessary to include several of the lower frequency modes to determine displacement. For simplicity the following sections consider the response in the principal mode which. the first higher mode involved in the vibration is the second symmetric mode which has a frequency considerably higher than the fundamental frequency.y) (Equation 8) where Φ(x. Consequently. The derivation of the response of a simply-supported floor to dance-type loads is given by Ji and Ellis[1] and extended to cover other structures. although for acceleration it may be sufficient to consider just one mode excited at resonance. In this Digest the derivation is not given and only the relevant equations are quoted.

0 + D) (Equation 18) β= fp fs Equations 11 and 12 give the steady state response which excludes transient terms. damping ratio ζ and mode shape F(x. Equations 11 and 12 can be expressed in the following form to accommodate the measurements. modal stiffness and mode shape. It is also preferable to use the measured damping value. At resonance. Alternatively.y) are required.y) dxdy s k* D = ∑ Dn n =1 Dn = √(1 – n2 β2)2 + (2nς β)2 rn. Equations 9 and 10 can be used where the floor mass density needs to be estimated. On the other hand.v sin (2πnft – θn + φn) (Equation 13) A= (1. If one person with a weight of Gs jumps at the centre of a floor the response is simply: A= Gs k* Ä= Gs (2πfs)2 k* Da (Equation 19) (1.y) = G) The structural displacement and acceleration can be expressed as: A=B G (1. .10 loading are uniformly distributed (ie m(x. it is desirable to perform dynamic tests.0 + D) m (2πf)2 G a D m (Equation 11) Using dynamic measurements in the analysis Ä=B (Equation 12) D and Da are defined as dynamic magnification factors for the displacement and acceleration. the acceleration in Equation 14 will be dominated by the term that occurs at the natural frequency of the structure. this is because the transient response decays quickly in a damped system and generally is of little interest[13].0 + D) (Equation 16) a Da = ∑ D n n=1 I Da = n2 β2 Dn n if 1 – n2 β2 >0 if 1 – n β =0 2 2 (Equation 14) Ä= G(2πfs)2 ∫∫ Φ(x. 3 For checking the safety and serviceability of existing structures or improving the accuracy of theoretical predictions. The fundamental frequency of most structures is relatively easy and inexpensive to measure[14]. G∫∫ Φ(x.y) dxdy s k* Da (Equation 17) π θn = 2    2nςβ  tan–1  1 – n2 β2 2nςβ  tan–1 +π  1 – n2 β2 if 1 – n2 β2 <0    (Equation 15) To use the above equations the measured modal stiffness k*. natural frequency f. it may prove difficult or inconvenient to measure the actual response when a crowd of people is involved.y) = m and G(x. damping. This will provide feedback from the actual structure including accurate values of the fundamental frequency.

The structural frequency above which vertical vibration should not pose a safety problem for jumping on floors is 8.4 Hz (ie 3 x 2. Therefore. If the calculations show an unacceptable response. below this level the vibration levels will certainly be felt by anyone remaining stationary on the structure. however. It is also likely that visible motion of a structure and items fixed to it would concern those immediately below the structure. even with knowledge of the source of the vibrations. related to perception of vibrations provide frequency weighting recommendations for motion in various directions. or to limit the crowd size or crowd activities. However. recommended peak accelerations for low frequency vibration are given in Table 3. for repetitive jumping. and. it is possible to define minimum frequencies for structures which should avoid safety problems from dynamic loads. The authors experienced vibrations of 15% g at 4 Hz on one structure and this certainly felt uncomfortable. avoiding resonance from the first three Fourier components of the load). then the likely design solution will be to increase the frequencies towards the given limits. and it is known that the frequency of the vibration is important. it is the acceleration levels which should be determined. for some structures. therefore the UK Standards. 35% g is a significant limit as above this level panic may occur. a frequency of 6 Hz[2] has been given as an interim measure (primarily avoiding resonance from the first two Fourier components of the load). calculation of the actual response will be appropriate. BS 6472 and BS 6841. From the Table it can be seen that. This may provide simple criteria which are convenient for the design of many structures. Ellis and Littler[16] explain how the acceleration levels given in the table can be linked to the VDVs suggested in the British Standards. a vibration level with a duration of one cycle (ie impulsive). however. These British Standards therefore recommend using vibration dose values (VDVs) which take account of these factors. Solutions which seek to increase the damping may have some attraction. it does not mean that all structures with lower frequencies will have problems. will have a different effect to a vibration level which is continuous. A number of experiments have been undertaken in Germany[15] on cantilevered grandstands. Table 3 Reaction to various peak acceleration levels on grandstands[15] (for a frequency range <10 Hz) Vibration level Reaction <5% g <18% g <35% g >35% g Reasonable limit for passive persons Disturbing Unacceptable Probably causing panic . the question of what are acceptable vibration levels must then be considered.11 Simple measures to avoid safety problems As dynamic crowd loading is confined to a narrow frequency range. these frequencies are for the appropriate mode of vibration of an empty structure (ie without a crowd)[7].8 Hz. A lot of work has been conducted on what vibration levels can be perceived. In addition. it seems appropriate to examine data obtained at such events. However. the number of cycles of the vibration is important when assessing acceptability of such vibrations. Serviceability limits For serviceability assessments related to human perception of vibrations. For guidance on acceptable vibration levels in structures subject to jumping loads. As mentioned previously. The frequency weighting effectively defines a filter that can be applied to the acceleration to compensate for the finding that human perception of vibration varies with frequency. For grandstands where concerts may be held. and as it is only necessary to consider a limited number of Fourier terms to evaluate displacements. they recognise that resonance will occur and attempt to limit the resonant response. and show that they provide a reasonable correlation with the observed audience reaction and measured response on a number of grandstands at a range of events.

Chapter 5. Loads due to spectator movements. The response of cantilever grandstands to crowd loads. McGraw-Hill. Code of practice for dead and imposed loads. 79 (6) 27–31. The Structural Engineer (18 November 2003). and the prevention and control of fire. Building Research Establishment. Glasgow.uk Details of BRE publications are available from: www. etc. (Accepted for publication in The Structural Engineer. Garston.com Published by BRE Bookshop Requests to copy any part of this publication should be made to: BRE Bookshop. Measured phase shifts in the dynamic response of a large permanent cantilever grandstand to human loading. Part 1: Serviceability evaluation. Derwent J M and Ji T. Code of practice for dead and imposed loads BS 6472:1992 Guide to evaluation of human exposure to vibration in buildings (1 Hz to 80 Hz) BS 6841:1987 Guide to measurement and evaluation of human exposure to whole-body mechanical vibration and repeated shock CP3:Chapter V:Part 1:1967. [16] Ellis B R and Littler J D. Two more ‘wobbly’ stands.co. Construction News.brebookshop. designers.) [12] Ellis B R and Littler J D. [10] Ji T and Wang D. Accepted for publication in the Procs of the Institution of Civil Engineers. we can take no responsibility for the subsequent use of this information. 241 (5) 920–924. [3] Rogers D. 17 August 2000. 1996. Dynamic performance requirements for permanent grandstands subject to crowd action – interim guidance for assessment and design. [7] Ellis B R and Ji T. Structures & Buildings (February 1997). 1975. [9] Tuan C Y and Saul W E. Munich. [13] Clough R W and Penzien J. [2] The Institution of Structural Engineers. 72 (3) 37–44. The Structural Engineer (1 February 1994). Human–structure interaction in vertical vibrations. Journal of Sound and Vibration (2001).com © Copyright BRE 2004 August 2004 ISBN 1 86081 715 7 Acknowledgements The authors would like to acknowledge the support of the Office of the Deputy Prime Minister (for the work undertaken by BRE) and the Engineering and Physical Sciences Research Council (for the work undertaken by University of Manchester Institute of Science and Technology) www. or for technical advice. A supplementary condition for calculating periodical vibrations. Actual problems with stand structures due to spectator-induced vibrations. [5] Ginty D. However. EURODYN’96. Watford WD25 9XX Tel: 01923 664761 Fax: 01923 662477 email: brebookshop@emap. 1991.bre.com or IHS Rapidoc (BRE Bookshop) Willoughby Road Bracknell RG12 8DW Tel: 01344 404407 Fax: 01344 714440 email: brebookshop@ihsrapidoc.co. [6] Littler J D.12 References [1] Ji T and Ellis B R.uk . December 2001. Procs of the Institution of Civil Engineers. [15] Kasperski M. 82 (6) 27–35. manufacturers. The frequency range and distribution of dance type loads. Accepted for publication in the Procs of the Institution of Civil Engineers. Dead and imposed loads BRE is committed to providing impartial and authoritative information on all aspects of the built environment for clients.bre. BRE IP 4/02 Loads generated by jumping crowds: experimental assessment British Standards Institution BS 6399:Part 1:1984. Loading for buildings. nor for any errors or omissions it may contain. 122 (1) 1–9. September 2002. Dynamic testing. BRE is the UK’s leading centre of expertise on building and construction. Florence. 5th European Conference on Structural Dynamics. Loading. We make every effort to ensure the accuracy and quality of information and guidance when it is first published. [4] Littler J D. Watford WD25 9XX Tel: 01923 664000 Fax: 01923 664098 email: enquiries@bre. Floor vibration induced by dance type loads – verification. Frequencies of synchronised human loading for jumping and stamping. [14] Ellis B R. Structures & Buildings. The response of cantilever grandstands to crowd loads. Floor vibration induced by dance type loads – theory. Contact BRE for information about its services. at: BRE. Part 2: Load estimation. Loading for buildings. ASCE Journal of Structural Engineering. BS 6399-1:1996. Monitoring building structures (Ed J F A Moore). [11] Ellis B R and Ji T. (February 1985). occupants. 111 (2) 418–434. The Structural Engineer (20 March 2001). engineers.uk Website: www. Code of basic data for the design of buildings.co. Structures & Buildings. New York. [8] Ellis B R and Ji T. Loads generated by jumping crowds: numerical modelling. contractors. Blackie & Son Ltd. 72 (3) 45–50. Dynamics of structures. The Structural Engineer (1 February 1994).

Sign up to vote on this title
UsefulNot useful